I have an equation, and one part of it is under an 'overbrace'. The problem is the numbers under the brace are too big. When I only write the part with the overbrace the size is fine. Is there any way this could be fixed? Thanks a lot! Code and picture attached

\documentclass[a4paper, 10pt]{article}

$\sqrt[n]{n} =^{(1)} \sqrt[n]{n \cdot 1^{n-1}} =\sqrt[n]{n \cdot 
\smash[b]{\underbrace{1 \cdots 1}_{n-1~times}
}} <^{(2)} \frac{n+\overbrace{1+\ldots+1}^{n-1~ times}}{n}=\frac{n+(n-1)}{n}=\frac{2n-1}{n}=2-\frac{1}{n}$

  • 1
    This is coded as inline math, which is why the fraction is sized for text. Will this really be used within text, or will it be a display? Coding it as a display should fix it. Commented Oct 29, 2022 at 17:52

2 Answers 2


The problem is that the Knuth's macro \overbrace (used in Plain TeX) declares explicitly \displaystyle for its argument, see line 959 of the plain.tex file. LaTeX, unfortunately, takes this idea too.

Solution: you can use Unicode Math, because Unicode Math declares \overbrace in completely different and better way. Or, you must explicitly say \scriptstyle in the argument of \overbrace. But this is poor solution: when you decide to switch your formula from inline math $...$ to display $$...$$ then you have to remove this \scriptstyle command. Best solution: use Uncode Math (for example OpTeX uses Unicode Math, XeLaTeX or LuaLaTeX are able to use Unicode Math too).

  • Not sure if the issue you're discussing is what's actually plaguing the OP. I thought it was that $n-1$ times -- which is typeset at \scriptsize-- was wider than the horizontal curly brace right above it.
    – Mico
    Commented Oct 29, 2022 at 18:28
  • 1
    @Mico OP used inline math, so the fraction has numerator, denominator in \sriptstyle. OP refers, that the digits "1" is bigger in the nominator. Note the picture. Of course, OP has a few mistakes in the code and you corrected them in your answer, but the core of the question is: digits "1" are bigger when inline math is used.
    – wipet
    Commented Oct 29, 2022 at 18:38
  • Thanks. You're absolutely right, and it is my working assumption regarding the nature of the OP's issue that's in error. I'll delete my answer.
    – Mico
    Commented Oct 29, 2022 at 18:53
  • @wipet now that Mico has deleted their answer I can't see what the mistakes in my code are, do you perhaps remember what they were?
    – tcb93
    Commented Oct 29, 2022 at 19:34
  • 1
    @Mico -- As requested by wipet, please undelete your answer. Commented Oct 29, 2022 at 23:09

I suggest you employ the mathtools package (which, incidentally, loads the amsmath package automatically) and use its \mathclap macro.

Oh, and don't forget to insert a \vphantom directive to make up for the depth lost because of the \smash[b]{\sqrt{...}} wrapper.

enter image description here

\usepackage{mathtools} % for \mathclap macro


=^{(1)} \sqrt[n]{n \cdot 1^{n-1}} 
=\sqrt[n]{n \cdot \smash[b]{
   \underbrace{1\cdots 1}_{ \mathclap{\text{$n-1$ times}} }}} 
<^{(2)} \frac{n+\overbrace{1+\dots+1}^{\text{$n-1$ times}}}{n}


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .